Proc. Nati. Acad. Sci. USA Vol. 84, pp. 7778-7782, November 1987 Neurobiology Electrophysiological and autoradiographical evidence of V, receptors in the lateral septum of the rat brain (binding sites/brain slices// subtypes/synthetic structural analogues) M. RAGGENBASS, ELIANE TRIBOLLET, AND J. J. DREIFUSS* Department of Physiology, University Medical Center, 1211 Geneva 4, Switzerland Communicated by Berta Scharrer, July 20, 1987 (received for review May 5, 1987)

ABSTRACT Extracellular recordings were obtained from tricular nucleus or from the amygdala (9, 10). Vasopressin single located in the lateral septum, an area known to can be released in a calcium-dependent manner from rat receive a vasopressinergic innervation in the rat brain. Ap- septum in situ (11) and in vitro (12). In addition, binding sites proximately half of the neurons tested responded to 8-L- for vasopressin have been detected by light microscopic arginine vasopressin (AVP) by a marked increase in firing rate autoradiography in the lateral septum (13-16). Taken togeth- at concentrations >1 nM. The effect ofvasopressin was blocked er, these data suggest that vasopressin may function as a by synthetic structural analogues possessing antagonistic prop- peptide transmitter in the lateral septum. erties on peripheral vasopressin and oxytocin receptors. To corroborate this conjecture we have carried out extra- Oxytocin was much less potent than vasopressin in firing septal cellular single-unit recordings from slices of the rat brain and neurons, and a selective oxytocic agonist was totally ineffective. have observed that vasopressin increases the excitability of The action of vasopressin on neuronal firing was mimicked by neurons located in the lateral septum. Next, to characterize the vasopressor agonist [2-phenylalanine,8-ornithinelvasotocin the receptor type(s) mediating this stimulatory effect of but not by the selective antidiuretic agonist 1-deamino[8-D- vasopressin, we have tested various synthetic structural argininelvasopressin. In a parallel study, sites that bind analogues of vasopressin that possess receptor-specific, [31H]AVP at low concentration (1.5 nM) were found by in vitro agonistic properties in peripheral tissues (17). In a parallel autoradiography in the lateral septum. Adjacent sections were study, we determined the distribution of [3H]vasopressin and also incubated with 1.5 mM [3HJAVP and, in addition, with 100 [3H]oxytocin binding sites in the septum by in vitro light nM [2-phenylalanine,8-ornithine]vasotocin or 1-deamino[8-D- microscopic autoradiography and assessed the speci- arginine]vasopressin-i.e., the same compounds as those used ficity of these binding sites by performing displacement tests for the electrophysiological study. Results showed that the using the same structural analogues. vasopressor agonist, but not the antidiuretic agonist, displaced [3H]AVP, thus indicating that the vasopressin binding sites MATERIALS AND METHODS detected by autoradiography in the septum were V1 (vasopres- sor type) rather than V2 (antidiuretic type) receptors. Based on Electrophysiological Recordings. Experiments were done the electrophysiological evidence, we conclude that these re- using male adult rats (200-300 g) from a Sprague-Dawley- ceptors, when occupied, lead to increased firing oflateral septal derived strain. The animals were decapitated, their brains neurons. were excised, and a block of brain tissue containing the septum was prepared. Coronal slices, 300-400 um thick, The neurohormone vasopressin is synthesized in the hypo- were cut using a vibrating microtome. Slices were transferred thalamus and carried by axoplasmic transport to the to a thermoregulated (34-35°C) recording chamber and neurohypophysis, where the arrival of action potentials placed at the interface between a humidified oxygenated triggers its release from nerve terminals into the bloodstream atmosphere and a perfusion solution of 130 mM NaCI/5 mM (1). Vasopressin then acts upon various peripheral tissues KCI/20 mM NaHCO3/1.2 mM KH2PO4/10 mM glucose/i after its binding to membrane receptors. The hypertensive mM MgSO4/0.7 mM CaCl2. The solution was gassed with effect of vasopressin is mediated by V1 receptors located on 95% 02/5% C02, pH 7.35-7.45. Peptides were tested by vascular smooth muscle; in addition, V1 receptors have been dissolving them in the perfusion solution, which flowed at -2 detected in several other tissues, such as , where ml/min and could be exchanged in 2.5-3.0 min. Under vasopressin causes glycogenolysis (2). Binding of vasopres- microscopic guidance, extracellular single-unit recordings sin to V1 receptors stimulates Ca2+ mobilization through a were obtained from the lateral septum using glass micropi- process that involves phosphatidylinositol breakdown (3). pettes filled with 4 M NaCI and having tip dc resistances of The antidiuretic effect of vasopressin is mediated by V2 5-25 MW. Voltage signals were amplified, filtered, and receptors located in the that activate a membrane displayed on an oscilloscope. Records of single-unit activity adenylate cyclase (2). were gathered with a rate meter and monitored on chart. Furthermore, it has become clear that vasopressin is Light Microscopic Autoradiography. The brains of animals synthesized by neurons in hypothalamic and extrahypothal- from the same stock were rapidly removed and frozen in amic nuclei the axons of which project to numerous areas within the mammalian central nervous system (4, 5). In rats, Abbreviations: AVP, 8-L-arginine vasopressin; OXT, oxytocin; [Phe2,Orn']VT, [2-phenylalanine,8-ornithine]vasotocin; HO[Thr4, the lateral division of the septum is densely innervated by Gly7]OXT, [1-(L-2-hydroxy-3-mercaptopropionic acid), 4-threo- vasopressin-immunoreactive axons and axon terminals (6, 7). nine,7-glycine]OXT; deamino-DAVP, 1-deamino[8-D-arginine]vaso- Some of these axons originate in the bed nucleus of the stria pressin; dEt2Tyr(Me)DAVP, [1-(J3-mercapto-43,,B-diethylpropionic terminalis (8), whereas others may derive from the paraven- acid),2-0-methyltyrosine,8-D-arginine]vasopressin; d(CH2)5Tyr- (Me)AVP, [1-(3-mercapto-,8,,3-cyclopentamethylene propionic ac- id),2-0-methyltyrosine]arginine vasopressin. The publication costs of this article were defrayed in part by page charge *To whom reprint requests should be addressed at: Department of payment. This article must therefore be hereby marked "advertisement" Physiology, University Medical Center, 9, Av. de Champel, in accordance with 18 U.S.C. §1734 solely to indicate this fact. CH-1211 Geneva 4, Switzerland.

7778 Neurobiology: Raggenbass et al. Proc. Natl. Acad. Sci. USA 84 (1987) 7779

Table 1. Endocrine activities (international units per mg) of For V1/V2 receptor discrimination, one series ofsections was AVP, OXT, and agonists incubated with 1.5 nM [3H]AVP; adjacent sections were Vasopressor Antidiuretic incubated with the same concentration ofvasopressin and, in Peptide (V1) (V2) Oxytocic* addition, with 100 nM of either the V1 agonist [2-phenylala- nine,8-ornithine]vasotocin ([Phe2,Orn8]VT) or the V2 agonist AVPt 369 323 25.5 1-deamino[8-D-arginine]vasopressin [Phe2,Orn8]VTf 121 1.6 5.3 (deamino-DAVP). Deamino-DAVPt 0.4 1200 2.9 Incubation was followed by two washes in ice-cold incu- OXT§ 4.3 4 486 bation medium and a quick rinse in distilled water. The slides HO[Thr4,Gly7]OXT§ <0.01 0.004 1002 were then dried with cold air, put for 2 hr in a vacuum des- sicator containing paraformaldehyde powder preheated at *Tested on rat uterus in vitro in the presence of 0.5 mM Mg2+. and then an tManning et al. (20). 80°C, placed in x-ray cassette in contact with *Berde et al. (21). tritium-sensitive LKB Ultrofilm for 3-4 mo at 4°C. §Lowbridge et al. (22). Chemicals. AVP and OXT were purchased from Bachem Fine Chemicals (Bubendorf, Switzerland). The synthetic isopentane at -250C. Series of 10- to 15-,um-thick sections structural analogues [Phe2,0M8]VT; deamino-DAVP; HO- were cut, mounted on gelatin-coated slides, dried for 2-12 hr [Thr4,Gly7]OXT; [1-(f-mercapto-f,3B3-diethylpropionic acid),- at 40C, and stored at -80'C until use. Control sections of 2-0-methyltyrosine,8-D-arginine]vasopressin [dEt2Tyr(Me)- kidney were also prepared. Sections were preincubated for 20 DAVP]; and [1-(,3-mercapto-fB3-cyclopentamethylene propi- min in 50 mM Tris HCl (pH 7.4) containing 100 mM NaCl and onicacid),2-0-methyltyrosine]arginine vasopressin [d(CH2)5- 50 nM guanosine 5-triphosphate, then rinsed in 50 mM Tyr(Me)DAVP] were provided by M. M. Manning (Depart- Tris-HCl. Incubation was then carried out for 1 hr in a humid ment of Biochemistry, Medical College of Ohio, Toledo, chamber by covering each section with 100-200 ,ul ofmedium OH). Three of these compounds are agonists possessing of 50 mM Tris-HCl/O.1 mM bacitracin/5 mM MgCI2/0.1% enhanced receptor specificity, as is shown in Table 1. The last bovine serum albumin containing various labeled and unla- two compounds are powerful V1 and oxytocic antagonists beled peptides. and possess extremely weak V2 agonistic activity (18, 19). For differentiation of oxytocin (OXT) and 8-L-arginine For autoradiography, AVP tritiated on the phenylalanyl vasopressin (AVP) binding sites, one series of sections was residue (specific activity, 40 Ci/mmol; 1 Ci = 37 GBq) and incubated with 1.5 nM [3HJAVP, and another series was tritiated OXT labeled on the tyrosyl residue (specific activity, incubated with 3.5 nM [3H]OXT. Adjacent sections were 35 Ci/mmol) were purchased from New England Nuclear. incubated with the same concentration ofone ofthese labeled Their radiochemical purity was checked by HPLC, the peptides and, in addition, with 150 nM of the specific fractions corresponding to the tritiated peptide being sepa- oxytocic agonist [1-(L-2-hydroxy-3-mercaptopropionic ac- rated on an affinity-chromatography column containing neu- id),4-threonine,7-glycine]oxytocin (HO[Thr4,Gly7JOXT). rophysin bound to Sepharose. AVP 5 10-8 M

d Et2Tyr(Me)DAVP 5 10-7M 9 AVP 5 10-8 M

20 29 AVP 5-10-8M

68 77 AVP 5 10-8M

-20 ] spikes/s

1 I I I I i 83 -. 91 min FIG. 1. Effects of AVP and of the synthetic structural analogue dEt2Tyr(Me)DAVP, on a septal . In this and the two following illustrations, the records are consecutive, but not continuous. AVP was added to the perifusion solution at the concentration indicated and for the period of time shown by a solid horizontal line above each record. The synthetic analogue, present at 500 nM for 6.5 min, starting from min 17 completely and reversibly suppressed the AVP-induced increase in firing. 7780 Neurobiology: Raggenbass et al. Proc. Natl. Acad. Sci. USA 84 (1987) RESULTS antidiuretic agonist, deamino-DAVP. These tests were done on nine neurons. The V2 agonist was without effect even Electrophysiology. Stable unitary extracellular recordings when applied at a high concentration; in contrast, all nine were obtained from over 100 neurons located in the dorsal neurons tested were readily excited by the V1 agonist: at two and intermediate areas of the lateral septum. They were to five times higher concentrations [Phe2,Orn8]VT produced either silent or fired irregularly. Vasopressin, at concentra- an effect comparable to that of AVP (Fig. 3). tions >1 nM, excited 52 neurons; 45 neurons did not respond, Autoradiography. In coronal sections that included the and 8 neurons were inhibited. The vasopressin-induced septum, the distribution of vasopressin and OXT binding excitation was concentration-dependent and fully reversible. sites differed markedly. Thus, whereas the septum contained This increase in firing rate was probably caused by membrane no binding sites for OXT (Fig. 4A), the lateral septum was one depolarization because at concentrations >100 nM, the of areas of the brain in which PHivasopressin labeling was excitation brought about by vasopressin could lead to partial most intense (Fig. 4C). OXT binding was detected in the action potential inactivation. Evidence that the action of the stria terminalis, and this vasopressin was receptor-mediated was provided by testing dorsal part of the bed nucleus of 15 neurons without and with the synthetic structural ana- binding could be displaced by HO[Thr4,Gly7]0XT (Fig. 4B). logues, dEt2Tyr(Me)DAVP and d(CH2)5Tyr(Me)AVP. These In contrast, HO[Thr4,GlyI]OXT could not successfully com- compounds did not by themselves affect the spontaneous pete with vasopressin binding (Fig. 4D). firing of the cells but fully and reversibly antagonized the In two more animals, the V1 character of [3H]AVP binding excitation induced by vasopressin (Fig. 1). in septum could be ascertained by using V1 and V2 agonists In order to assess which receptors were involved, the in competition studies. Results are shown in Fig. 5, with effects of vasopressin, of OXT, and of the selective oxytocic kidney slices included for comparison. In contrast to renal agonist were tested on eight neurons, which were all excited labeling, vasopressin binding in septum was displaced by the by vasopressin. OXT also had an excitatory action, but it was V1 agonist (Fig. 5D) but not by the V2 agonist (Fig. 5H). less potent by a factor of 10-200 than vasopressin (Fig. 2, third trace). The selective oxytocic agonist had no effect, DISCUSSION even when applied at a high concentration (Fig. 2, fourth trace). We have shown that the predominant effect ofvasopressin on To further characterize the vasopressin-induced response, neurons located in the lateral septum of rats is excitatory; the action of vasopressin was compared with that of a V1 moreover, the effects of receptor specific agonists indicated agonist, [Phe2,Orn8]VT, and with that of the selective that this action of vasopressin was mediated by V1 receptors. AVP 5 10-8M

IIIIIIII 0 8 OT 5 10-8M

10 18 OT 5 10-7M

20 28 HO [Thr4, Gly7] OT 5 10-7M

30 38 AVP 510-8M

spikes/s I I 39 46 min FIG. 2. Effects ofAVP, ofOXT (OT), and ofthe selective oxytocic agonist HO[Thr4,Gly7]OXT (OT) on the firing activity of a septal neuron. Note that the selective oxytocic agonist had no effect2~~~~~~~~~~~~~~~~~and that OXT was weaker by a factor of >10 than AVP in exciting the cell. Neurobiology: Raggenbass et al. Proc. Natl. Acad. Sci. USA 84 (1987) 7781 AVP 5 10-8M a V1 agonist, but neither a V2 nor an oxytocic agonist, displaced the labeled compound. Taken together, our results corroborate the notion that endogenous vasopressin might act as a peptide transmitter in the lateral septum. Earlier reports had suggested that vasopressin and OXT exert an excitatory effect on a small proportion of neurons 0 7 located in the lateral septum (23, 24). In addition, both peptides were said to enhance the glutamate-induced exci- [Phe2.Orn8] VT2.107M tation of septal neurons and to increase the number of action potentials evoked by orthodromic activation ofthese neurons (24, 25). In contrast, however, other workers reported that, in the same brain area, vasopressin suppressed synaptically evoked action potentials (26). In these studies, the neuro- peptides were administered locally by microiontophoresis, IIi II and their effective concentration was therefore unknown; 33 40 moreover, neither agonists nor antagonists were tested to dDAVP 10-6M ascertain the selectivity of the effects. More recently, Urban and De Wied (27) showed that vasopressin, superfused on the dorsal surface of the septum, could increase the negative wave ofa field potential evoked by stimulation ofthe fimbria.

L The results were somewhat surprising, however, because 52 59 OXT was as powerful as vasopressin in eliciting this effect. Previous work, using light microscopic autoradiography, [Phe2, Orn8] VT 2 10-7M has shown that the lateral septum of rats contains [3H]vaso- pressin binding sites (13-16). Biegon et al. (15) used displace- ment studies to report that d(CH2)5Tyr(Me)AVP, but not 20 OXT, competed efficiently with vasopressin for binding. ]ikes/sp_ These findings lead the authors to conclude, in accordance with the present study, that septal vasopressin binding sites are of the V1 type. In contrast to our results, De Kloet et al. 82 88 min (16) reported that [3H]OXT labeled the lateral septum, though FIG. 3. Effects of AVP and of synthetic structural analogues on less effectively than [3H]vasopressin. This discrepancy may a septal neuron. [Phe2,Orn8]VT, a V1 agonist, added to the perifusion be due, at least in part, to the fact that these authors used 10 solution at 200 nM, caused an increase in firing comparable to that nM of tritiated peptide, a value much higher than those used brought about by AVP at 50 nM. In contrast, deamino-DAVP in the present study. (dDAVP), 1000 nM, was without effect. High-affinity binding sites for vasopressin, with a Kd in the nanomolar range, have been characterized in at least two The autoradiographic data were in accordance with the laboratories on isolated membrane preparation. In the results of the electrophysiological studies. A high density of amygdala and in the dorsal hindbrain of rats, d(CH2)5Tyr- [3H]AVP binding sites was detected in the lateral septum, and (Me)AVP, but not OXT, was an effective competitor of competitive studies using unlabeled structural analogues [3H]vasopressin binding, a finding that indicated that the sites suggested that these binding sites were ofthe V1 type because present in these areas ofthe brain resemble V1 rather than V2,

A Cli. B

44.chTs

A-

9 T: !-; .' 0 e"

4

I,

A.. It BST

FIG. 4. Autoradiographs reveaiing the presence of binding sites for OXT and AVP in sections of the rat brain. Two adjacent sections were exposed to 3.5 nM [3H]OXT alone (A) or to 3.5 nM [3H]OXT and to 150 nM HO[Thr4,Gly7]OXT (B). In A, OXT binding sites are present in the bed nucleus of the stria terminalis (BST) and in the choroid plexuses that span the wall of the lateral ventricle (ChP); (B) binding in these sites was displaced by the selective oxytocic agonist. Note the absence of [3H]OXT binding in the lateral septum (A), and contrast it with an adjacent section, incubated with 1.5 nM [3H]AVP (C). [3H]AVP binding in septum and BST was not displaced by coapplication of 150 nM HO[Thr4,Gly7]OXT (D). Calibration line: 2 mm. 7782 Neurobiology: Raggenbass et al. Proc. Natl. Acad. Sci. USA 84 (1987) receives a sexually dimorphic dense vasopressinergic inner- vation appears to be particularly suited for studies aimed at * better understanding the role(s) of vasopressin in the central nervous system. We thank Dr. M. M. Manning (Toledo, OH) for the supply of peptides and Drs. C. Barberis and S. Jard (Montpellier, France) for valuable discussions and for help in peptide purification. This work was supported in part by Grant 3.358.086 from the Swiss National Science Foundation and the Fondation Carlos et Elsie de Reuter.

C.. / z ) _ 1. Poulain, D. A. & Wakerley, J. B. (1982) Neuroscience 7, 773-808. 2. Jard, S. (1983) in Current Topics in Membrane and Transport: 1 4~~~~~~~~~~~~~~~~~~~ 7'. Membrane Receptors, ed. Kleinzeller, A. (Academic, New York), Vol. 18, pp. 255-285. 3. Michell, R. H., Kirk, C. J. & Billah, M. M. (1979) Biochem. Soc. Trans. 7, 861-865. 4. Sofroniew, M. V. (1985) in Handbook ofChemical Neuroanat- omy: GABA and in the CNS, eds. Bjorklund, A. & Hokfelt, T. (Elsevier, Amsterdam), Vol. 4, pp. 93-165. 5. Muhlethaler, M., Raggenbass, M. & Dreifuss, J. J. (1985) in F , . .. s Actions in the Vertebrate Nervous System,

p eds. Rogawski, M. A. & Barker, J. L. (Plenum, New York), _ , ~> pp. 439-458. 6. De Vries, G. J., Buijs, R. M., Van Leeuwen, F. W., Caffd, A. R. & Swaab, D. F. (1985) J. Comp. Neurol. 233, 236-254. 7. Buijs, R. M. & Swaab, D. F. (1979) Cell Tissue Res. 204, 355-365. 8. De Vries, G. J. & Buijs, R. M. (1983) Brain Res. 273, 307-317. 9. Hoorneman, E. M. D. & Buijs, R. M. (1982) Brain Res. 243, 6 H 235-241. Jilft '1111111111lik 10. Caffe, A. R., Van Leeuwen, F. W. & Luiten, P. G. M. (1986) Neurosci. Abstr. 12, 299. 11. Demotes-Mainard, J., Chauveau, J., Rodriguez, F., Vincent, .. , .114. z; 0..z " -.V ." J. D. & Poulain, D. A. (1986) Brain Res. 381, 314-321. 12. Buijs, R. M. & Van Heerikhuize, J. J. (1982) Brain Res. 252, 71-76. 13. Van Leeuwen, F. W. & Wolters, P. (1983) Neurosci. Lett. 41, 61-66. 14. Baskin, D. G., Petracca, F. & Dorsa, D. M. (1983) Eur. J. FIG. 5. AVP binding sites in lateral septum and their pharmaco- Pharmacol. 90, 155-157. logical identification as VI receptors. Sections from kidney (A, B, E, 15. Biegon, A., Terlou, M., Voorhuis, T. D. & De Kloet, E. R. and F) and brain (C, D, G, and H) were incubated with 1.5 nM (1984) Neurosci. Lett. 44, 229-234. [3H]AVP. B and D were incubated with the same amount of tritiated 16. De Kloet, E. R., Rotteveel, F., Voorhuis, T. A. M. & Terlou, AVP and in addition with 100 nM of [Phe2,Orn8]VT. This VI agonist M. (1985) Eur. J. Pharmacol. 110, 113-119. displaced label in septum (D) but not in kidney (B). Sections F and 17. Manning, M. & Sawyer, W. H. (1985) in Vasopressin, ed. H were exposed to 1.5 nM [3H]AVP and to 100 nM deamino-DAVP. Schrier, R. W. (Raven, New York), pp. 131-144. Note that the V2 agonist competed with AVP in kidney (F) but not 18. Manning, M., Lammek, B., Bankowski, K., Seto, J. & Saw- in septum (H). Calibration lines: 2 mm. yer, W. H. (1985) J. Med. Chem. 28, 1485-1491. 19. Kruszynski, M., Lammek, B., Manning, M., Seto, J., Haldar, or OXT receptors (28, 29). In a very thorough study in which J. & Sawyer, W. H. (1980) J. Med. Chem. 23, 364-368. synaptic membranes from rat hippocampus were used and 20. Manning, M., Balaspiri, L. & Moehring, J. (1976) J. Med. the binding of nanomolar concentrations oflabeled vasopres- Chem. 19, 842-845. sin, labeled OXT, and various synthetic structural analogues 21. Berde, B., Boissonnas, R. A., Huguenin, R. L. & Sturmer, E. was studied, Audigier and Barberis (30) reported the exist- (1964) Experientia 20, 42-43. ence of two distinct populations of binding sites, which 22. Lowbridge, J., Manning, M., Haldar, J. & Sawyer, W. H. probably correspond to V1 and OXT receptors, respectively. (1977) J. Med. Chem. 20, 120-123. 23. Huwyler, T. & Felix, D. (1980) Brain Res. 195, 187-195. To our knowledge, no comparable studies have yet been done 24. Joels, M. & Urban, I. J. A. (1982) Neurosci. Lett. 33, 79-84. using membranes isolated from the septum. 25. Joels, M. & Urban, I. J. A. (1984) Brain Res. 311, 201-209. The autoradiographic and electrophysiological data report- 26. Marchand, J. E. & Hagino, N. (1982) Exp. Neurol. 78, ed in the present article show convincingly the existence in 790-795. the rat septum of V1 receptors similar to those present on 27. Urban, I. J. A. & De Wied, D. (1986) Neuropeptides 7, 41-49. 28. Dorsa, D. M., Petracca, F. M., Baskin, D. G. & Cornett, smooth muscular, hepatic, and other nerve cells but mark- L. E. (1984) J. Neuroscience 4, 1764-1770. edly different from V2 receptors found in the medulla of the 29. Cornett, L. E. & Dorsa, D. M. (1985) Peptides 6, 85-89. kidney. We conclude that this part of the brain, which 30. Audigier, S. & Barberis, C. (1985) EMBO J. 4, 1407-1412.